Since we agree that it is a wash, I stand by my opinion that this realization puts more importance on the availability of fast lenses when making a camera system choice(where low light shooting is a priority).
I understand that you feel Olympus is far below the others, but regardless of sensor size I have seen samples that appear to be usable in small prints and the OP feels it may actually yield slightly better results than the other camera being considered. Hence I suggested lens selection should be given consideration.
FWIW, I never said that Olympus is far below the others. From what I've seen, it certainly does suffer more from noise than DSLRs with APS-sized sensors, but it's hardly unusable and certainly way beyond any PnS.
From what I can tell Olympus makes 1 fast prime under $1000(50mm F/2.0=100mm 35mm eqiuv) and NO OTHER fast primes or zooms under $2000(per BHphoto site).
This I did not realize, in which case I would agree that this makes the Olympus line-up less attractive.
However, supposedly the 4/3rds mount is the most interchangable of any DSLR mount, allowing you to use all manner of other lenses on them - yes, in manual focus only, but that is still a valuable tool. The 2x crop is certainly going to hurt though, making the common and high-quality 50mm lenses be a longish 100mm - the 75mm on a 1.5x crop body is already plenty long for what they're used for.
I do find it amusing that we seem to be agreeing that we wouldn't recommend the Olympus line but for different reasons.
As for the other comment, I would just ask that you not think about what camera I use when it has no place in whatever thread we're working in. The Canon comment seemed like a shot as well as I'm not exactly a fan of theirs, and I certainly don't agree than they are "clearly" better at high ISOs than their competition. (It sounds like you were making that because Canon's sensors are slightly smaller, but I had
no idea that that was your point.) It's also not true that the K10D has poor jpg performance, it has a specifically-chosen jpg performance which was designed for a different look (more film-like and less artificially sharpened) than other cameras. Those settings can also be changed, which Phil apparently decided not to do, but instead to turn it into the DSLR version of the phantom Audi "unintended acceleration" of the '80s. (Just because it's repeated often doesn't make it true.) It's also rather silly as anyone with a camera as advanced as the K10D is doing themself a disservice by not shooting in Raw mode.
And as for DPReview's numerical ratings themselves, I don't think those are anything to go by. Something as intangible as image quality can't really be tossed in a number, and the ratings certainly aren't consistent - a Fuji F30 gets an 8.5 in IQ from them - does anyone really expect it to be match those DSLRs or be better than the A100 or K10D? And the Fuji S6000 with exactly the same sensor gets an 8.0 in IQ -

It's kind of like how Roger Ebert always hated having to give number ratings for movies, they were not a good way or judging whether or not a movie was any good. Similarly, cameras cannot be easily broken down into a specific numbers, either.